Skip to content

Intermittent Sampling

Sample periodically or irregularly to detect intermittent states that continuous monitoring cannot afford or guarantee.

The Diagnostic Story

Symptom: The problem is real — people report it, the data shows traces of it — but every direct inspection finds normal conditions. Evidence appears in bursts and then vanishes, leaving no consistent picture. Repeated checks always catch the same partial view, so the failure can neither be confirmed nor ruled out. Debate stalls on whether the phenomenon is genuine because proof is sporadic and disappearing.

Pivot: Accept that observation will be incomplete, then deliberately design when, where, and how samples are taken. Define the interval or trigger, the detection window, what counts as a positive, and what happens next when one is found.

Resolution: Intermittent conditions become observable enough to diagnose or govern. Continuous monitoring costs are avoided where full coverage was never achievable. As the design improves over iterations, the debate shifts from anecdotal claims to sampled evidence and positive detections drive earlier, defined responses.

Reach for this when you hear…

[clinical diagnostics] “The arrhythmia only shows on a 24-hour Holter, not on the 12-lead we do in the office -- you have to catch it when it is actually happening.”

[network operations] “Packet loss that is gone by the time I run a trace is still packet loss -- I need periodic captures at the right interval, not just a ping when someone complains.”

[quality inspection] “We cannot test every unit off the line, but if we sample at the right points in the shift we will catch the drift before it becomes a batch failure.”

Mechanisms / Implementations

  • Spot Checks: Implement short bounded observations.
  • Randomized Audits: Use unpredictability to reduce gaming and patterned blind spots.
  • Canary Probes: Are lightweight signals or tests placed in the system to reveal whether an intermittent state may be occurring.
  • Diagnostic Sampling: Collects targeted measurements when a condition is suspected but not continuously visible.
  • Burst Capture Logging: Increases observational density around suspected bursts.
  • Rotating Inspections: Sample different units, sites, teams, or subsystems over time.
  • Sentinel Surveys and Temporary Sensors: Sentinel surveys gather evidence from selected people, locations, or institutions that are likely to show early signal.
  • Canary Probe
  • Randomized Audit
  • Rotating Inspection

Abstractions this archetype builds on — directly (a source ingredient) or as a related pattern. Links follow the typed catalog namespace.

Built directly on (3)

Also references 14 related abstractions

Variants

Narrower or domain-specific specializations that share this archetype's core structure. Recognized variants are established; candidate variants are provisional.

Scheduled Intermittent Sampling · temporal variant · recognized

Sampling at planned intervals to make intermittent states visible without continuous monitoring.

Randomized Intermittent Sampling · implementation variant · recognized

Sampling at unpredictable times or cases to reduce bias, gaming, and patterned misses.

Triggered Intermittent Sampling · temporal variant · recognized

Opening a sample window when a proxy signal suggests the intermittent state is likely to appear.

Sentinel Sampling · scale variant · recognized

Sampling selected sentinel locations, actors, devices, or cases that are likely to reveal intermittent states earlier or more cheaply than broad coverage.

Burst-Window Sampling · risk or failure variant · candidate

Sampling densely during suspected burst windows and sparsely outside them.